> ## Documentation Index
> Fetch the complete documentation index at: https://api.thesapientcompany.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Use cases & applications

> Real ways teams use Sapient, and which model fits each.

Sapient reads how a real human brain responds to your content, second by second. Below are the most common ways teams put that to work, in plain language: what you're trying to learn and what to submit. Every case runs on [Qualia](/api-reference/models), the active model.

<Note>
  **Qualia is the model for every case below.** It's the default and only available model right now — submit a video or audio URL and omit the `model` field. Mary (which read scripts, copy, and other text) is **temporarily unavailable**, so text input isn't currently supported.
</Note>

## A/B testing ad creatives

Compare two ad variants and see which one actually holds attention and drives purchase intent, before you decide where the budget goes.

* **Submit:** each video variant as its own scan, then compare the timelines.
* **Model:** **Qualia** — it's sharpest exactly here, separating visually distinct variants (different cuts, footage, pacing). (Script- or copy-only comparisons used Mary, which is temporarily unavailable.)
* **Look at:** the [attention](/knowledge-base/lenses/attention) and [purchase intent](/knowledge-base/lenses/purchase-intent) curves side by side, to see which one stays higher, longer.

## Hook analysis & optimization

The first three seconds decide whether anyone keeps watching. Find exactly where the hook lands and where attention drops off.

* **Submit:** the creative (reel, TikTok, ad) as a video URL. **Model:** **Qualia**.
* **Look at:** the per-second [attention](/knowledge-base/lenses/attention) curve over seconds 0–3, plus the `moments` block for the first flagged drop.

## Creator content pre-post check

Score a reel or TikTok *before* you publish it, so you ship the version that performs.

* **Submit:** the finished cut as a video URL. **Model:** **Qualia** (it reads the visuals and audio together).
* **Look at:** the overall score and grade for a fast verdict, then the timeline to tighten weak seconds.

## Pre-flight QA before ad spend

Before you put real budget behind a creative, sanity-check it the way a focus group would, instantly.

* **Submit:** the creative as a video URL. **Model:** **Qualia** — especially strong when you're choosing between visual variants.
* **Look at:** the headline score, the [attention](/knowledge-base/lenses/attention) and [purchase intent](/knowledge-base/lenses/purchase-intent) lenses, and the `moments` where it slips.

## Voiceover testing

Compare voiceover variants line by line, without filming anything.

* **Submit:** the voiceover as an `audio` URL. **Model:** **Qualia**.
* **Look at:** the per-second [attention](/knowledge-base/lenses/attention) and [purchase intent](/knowledge-base/lenses/purchase-intent) curves to see which lines land and which lose the listener.

<Note>
  Scoring a raw **script or ad copy as text** was Mary's specialty, and Mary is **temporarily unavailable** — text input isn't currently supported. Test the voiceover as audio in the meantime.
</Note>

## Manipulation & brand-safety check

See how heavily a creative leans on emotional persuasion versus plain reasoning, which is useful for brand-safety and compliance review.

* **Submit:** the ad as a video or audio URL. **Model:** **Qualia**.
* **Look at:** the [manipulation](/knowledge-base/lenses/manipulation) lens across the timeline.

## Memorability test

Attention is cheap; memory is what sells. Find out what actually sticks.

* **Submit:** the creative as a video or audio URL. **Model:** **Qualia**.
* **Look at:** the [memory](/knowledge-base/lenses/memory) lens. A high attention score with a low memory score means it grabs focus but doesn't stay.

<CardGroup cols={2}>
  <Card title="Models" href="/api-reference/models">
    Qualia, the active model — what it reads and why Mary is temporarily unavailable.
  </Card>

  <Card title="The 4 lenses" href="/api-reference/lenses">
    Every lens you can score a scan against.
  </Card>
</CardGroup>